v1.1.2

Additions:

• Using broadcast_to() simplify mask_static operations
• Fix: .detach() before .cpu()
• Fix: handle case where mask_static is not a torch.Tensor

v1.1.1

Additions:

• Compatible with PyTorch tensor as inputs
• Keeping dtype from PyTorch tensor or numpy array

v1.1

Additions:

• Based the TV direct calls and subgradients constructions on the tv_operators.
• Simplified the tv_operators implementation.
• Added time regularization and mask_static options for all functions, both CPU and GPU, all discretizations. Now everything is fully 4D.

v1.0

First release